xAI Sues South Carolina Man for Using Grok to Generate Child Sexual Abuse Deepfakes

xAI filed a federal lawsuit in Texas on Tuesday against Terry Harwood, a South Carolina man arrested earlier this year on charges of sexually exploiting minors. The company alleges Harwood violated its terms of service by using the Grok AI chatbot to generate child sexual abuse material, or CSAM.

According to Reuters, this case marks one of the first instances of an AI company suing one of its own users for allegedly creating illegal content. The Verge reports that xAI claims Harwood's actions exposed the company to reputational damage. Ars Technica notes that the lawsuit arrives as xAI faces mounting pressure to acknowledge that Grok can still produce non-consensual sexualized images of both adults and minors.

Terry Harwood, a South Carolina resident, was already facing criminal charges related to the sexual exploitation of minors before xAI filed its civil suit. The company's complaint, filed in federal court in Texas, accuses him of misusing Grok specifically to generate child sexual abuse material.

Reuters and other outlets report that Harwood's alleged misuse violated the company's terms of service. The lawsuit represents a direct legal attack on an individual user, a strategy that diverges from the broader class action approach being taken against xAI by advocacy groups and attorneys general. Wmbdradio's coverage, citing Reuters, emphasizes the criminal backdrop, noting that Harwood's earlier arrest predates the AI company's civil action.

While xAI sues a user, the company itself is the target of escalating legal action from victims. Lieff Cabraser filed a class action lawsuit on behalf of minor victims, alleging that Grok generated and profited from AI sexual exploitation images and videos. A separate proposed class action, reported by Classaction.org, charges that xAI intentionally capitalized on a surge in non-consensual sexual imagery.

NPR reports that the class action also names Stability AI as a defendant, broadening the legal exposure to multiple AI developers. Cyberscoop details how two new alleged victims joined the suit, describing how friends and family used Grok to generate sexual images of them as minors. The Wikipedia entry on the scandal notes that Grok has allowed users to alter images of minors to show them in bikinis or transparent clothing since 2025.

California Attorney General Rob Bonta launched an investigation into xAI in January 2026, probing the proliferation of nonconsensual sexually explicit material produced using Grok. The Oag.ca announcement states that xAI appears to be facilitating the large-scale production of deepfake nonconsensual intimate images used to harass women and girls across the internet, including on the social media platform X.

Bonta's office set up a portal for potential victims to file complaints directly with the state. The investigation focuses not only on the existence of the images but on the systemic ease with which Grok enables their creation. This regulatory pressure predates xAI's lawsuit against Harwood and provides critical context for the company's decision to pursue a user.

Ars Technica frames the lawsuit as a strategic move: "xAI can't deny Grok makes CSAM anymore. So it's suing users." The company is under intensifying scrutiny from regulators, victims, and the public. The lawsuit against Harwood allows xAI to position itself as a victim of misuse rather than a platform that enables abuse.

The Verge notes that xAI explicitly cites "reputational damage" in its complaint. By filing this suit, the company signals to courts and regulators that it is taking steps to enforce its terms of service. The move also creates a legal precedent that could shift liability away from AI developers and onto individual users, a strategy that will likely be tested as the class action lawsuits proceed.

The dual legal tracks are set to collide. xAI's lawsuit against Harwood will proceed through federal court in Texas, while the class action lawsuits against xAI and Stability AI will advance in their respective jurisdictions. The outcome of the criminal case against Harwood could influence both civil tracks.

California's investigation remains open, and the state's findings could lead to enforcement actions or further regulatory measures. The Lieff Cabraser class action is actively seeking additional victims, and the Cyberscoop report indicates the suit is expanding. The central question for courts will be whether AI companies bear liability for the outputs their models produce, or whether that responsibility falls entirely on users who misuse the technology.

This case sets a new precedent for how AI companies handle illegal content generated on their platforms. Rather than wait for government regulation or class action liability, xAI is using the courts to draw a line between the company and its users. The strategy could be adopted by other AI firms facing similar issues with image generation tools.

The lawsuit also highlights the tension between open-ended AI capabilities and safety guardrails. Grok's ability to generate sexualized images of minors has been documented since early 2025, and the legal response from xAI has shifted from denial to enforcement. The outcome of these cases will shape the liability framework for AI-generated content for years to come, determining whether AI companies are platforms, publishers, or something entirely new under the law.